DNA-R1B1C7-L Archives

Archiver > DNA-R1B1C7 > 2008-01 > 1200582548


From: "David Wilson" <>
Subject: Re: [DNA-R1B1C7] Some Musings on R1b1c7
Date: Thu, 17 Jan 2008 07:09:08 -0800
References: <c8a.1b4339a3.34c043c0@aol.com>
In-Reply-To: <c8a.1b4339a3.34c043c0@aol.com>


I also feel uneasy about what we actually mean when we talk about the "age"
of R1b1c7. We can see that the R1b1c7 modal haplotype stands at a
considerable distance from the modal haplotype for mainstream R1b1c -- a
genetic distance of 18 over 76 markers (compare C7BED and M5UKQ at Ysearch).
To my mind that implies divergence a long time ago. Note that R1b1c9 and
R1b1c10 show enough variance within their own populations to imply founders
who lived at least a few thousand years ago. But for all the observed
variance, their modal haplotypes are still quite close to the R1b1c modal
haplotype. R1b1c7 shows a fair amount of variance in its own right, but with
a modal haplotype that is quite a bit further away from the R1b1c modal.

The picture that best explains this for me is an early population division
that isolated the group within which the R1b1c7 haplogroup would eventually
develop. After two or three hundred generations, with lots of mutations and
lots of attendant Y-line extinctions, we get to a point where drift has
created a distinctive Y-DNA pattern that looks a lot like modern R1b1c7 but
in which the M222 mutation has not yet occurred. I am thinking this could
have been about 2,000-3,000 years ago.

Then the M222 mutation occurs in a member of this group. His haplotype was
probably quite similar to the modal haplotype of the R1b1c7 population
today, though based on the evidence of a couple of SNP-confirmed R1b1c7s
that lie some distance from the R1b1c7 modal, maybe there was a significant
variation or two. At any rate, coming forward again we get to the individual
who really is the direct ancestor of the majority of R1b1c7s today. Conn?
Niall? Someone more recent? A name probably doesn't matter at this point as
long as we see the relative chronology.

After a long period of resistance, I am finally allowing myself to consider
the possibility that the majority of the variance we see in R1b1c7
haplotypes today is the product of the last thousand years. It is
interesting to me that distinctive allele values, when found, are usually
associated with individual families, but not with groups of families. With
rare exceptions, there are almost no close regional associations with
particular allele values. It is hard to spot candidate alleles that would
unite two families to the exclusion of all others, and in cases where a rare
allele value is seen in two different families, it is apparently the result
of independent mutations because each family's other distinctive allele
values are not shared. If I recall correctly, a few months ago someone
reported (John, was that you?) the star-like nature of a network analysis of
the available R1b1c7 haplotypes. There were no leggy structures or
pronounced branches to the picture -- there was a core area with elements
radiating in all directions, which is what we should expect only 30 or 40
(my suggested numbers, so don't blame anybody else) generations downstream
from a founder.

David Wilson



-----Original Message-----
From:
[mailto:] On Behalf Of
Sent: Wednesday, January 16, 2008 9:38 PM
To:
Subject: [DNA-R1B1C7] Some Musings on R1b1c7

The Trinity Study.

The Trinity study, based on a sample of 59 surnames linked by pedigree to
Nial 'of the Nine Hostages' (d. 405 A.D. ), arrived at two different TMRCA
estimates; one was 1730 years ago, or about 277 A.D. The second estimate
was 1,010 years ago (or about 997 A.D.). The team concluded the first
estimate was "at least consistent with an early medieval time frame." The
second suggested an origin predating the adoption of surnames in Ireland
(generally beween 950 A.D. and 1100 or later). Of these 59 samples roughly
half matched the IMH modal or what is now called R1b1c7. The Trinity team
suggested the non matching samples were the result of "the cumulative
consequence of nonpaternity events and the induction into the clan
structure of unrelated males. "

The Following surnames were involved in the 59 sample study.

(O')Gallagher (12), (O')Boyle (9), (O')Doherty (5), O'Donnell (4), O'
Connor
(3), Cannon (3), Bradley (2), O'Reilly (2), Flynn (2), (Mc)Kee (2),
Campbell (1), Devlin (1), Donnelly (1), Egan (1), Gormley (1), Hynes (1),
McCaul (1), McGovern (1), McLoughlin (1), McManus (1), McMenamin (1),
Molloy (1), O'Kane (1), O'Rourke (1), and Quinn (1).

Anyone familiar with Ui Neill surnames in Ireland will immediately recognize
many of these surnames. Most are either Cenel Conaill or Cenel Eoghain,
said to descend from two different sons of Nial 'of the Nine Hostages.' (d.
405
A.D.) Perhaps less well known is the fact that a group of these surnames
fall into the tribal designation of Connachta, who are not said to descend
from Nial in the pedigrees. Instead the pedigrees have them descend from
brothers of Nial, a genealogical statement few historians accept as
historically accurate. But the pedigrees do claim at least a genealogical
link between the Ui Neill (descendants of Nial) and the Connachta
(descendants of Conn, an earlier ancestor in the pedigree) and DNA appears
to validate the claim.

Nowhere in their article does the Trinity team actually state that Nial 'of
the Nine Hostages' was the common ancestor of R1b1c7 in Ireland. They
point only to an "early medieval time frame" more or less consistent with
the time of Nial (but actually predating him by several centuries).
Associating this common ancestor with Nial is a leap the Trinity team did
not make although nearly everyone else has.

It's also important to point out that the Trinity study included only Irish

samples. In their article, they cited an earlier Capelli Y chromosome
census
of the British Isles which featured a much reduced marker set and noted
that "The 6-STR IMH is virtually absent from much of Britain but reaches
frequencies of up to 7.3% (16.7% including likely one-step derivatives) in
western and central Scottish locations." The article did not mention that
the Capelli study also found a few similar haplotypes in Denmark and
Norway. Attempting to identify R1b1c7 from such a limited marker set is
problematic at best.

The Trinity study therefore did not address the question of when R1b1c7
originated. It simply estimated the TRMCA of a sub-group called the Ui
Neill in Ireland (to which for accuracy's sake we must add the Connachta).
They made no attempt to to factor in Scottish or English R1b1c7 samples
although they were aware it existed.

Most of the leading DNA experts on the GEN-DNA list believe R1b1c7
originated in Ireland in the northwest. Or if not in the northwest at
least in
Ireland. Estimates vary as to the TMRCA but John McEwen for one has
stated it
could have been many thousands of years ago, as early as 3,000 BC. (which
he calls a conservative estimate). I have no idea what dataset John McEwen
used in his estimate but I'm sure it included a lot of Scottish samples
(possibly his
R1bSTR19 cluster).

At a glance there seems to be quite a dichotomy between the estimates on
GEN-DNA and the Trinty estimates of 275 A.D. But perhaps they were looking
at
different aspects of the same puzzle. The two estimates might not be so
incompatible after all.

Is this a possible explanation for the wildly varying dates given in the
Trinity study versus the opinions of DNA experts? That the Trinity study
considered only DNA samples from surnames alleged to be descendants of Nial
or the Connachta in Ireland and other estimates include the whole range of
R1b1c7 from Ireland to Scotland? Would adding Scottish and English DNA to
the surname mix used by Trinity result in a different TMRCA for R1b1c7?

It might be an interesting experiment but I don't know how to conduct such
a test. The Trinity team stated they used Network.exe to estimate TMRCA
for their surnames. How did John McEwen arrive at his own estimates?

"The time to the most-recent common
ancestor (TMRCA) of this lineage was estimated with the r statistic (Morral
et al. 1994) in NETWORK, with use of a mutation rate of 1 per 2,131 years
for a 17-marker haplotype (Zhivotovsky et al. 2004)."

Under Time Estimates in Network, the following parameters are asked:

1 mutation every

20,180 [default] years

Then you are asked to select an ancestral node. What do you chose here?
The R1b1c7 modal? Or will any large node do?

Then you have to select descendant nodes.

To use this I assume you'd have to come up with a years figure based on 25
or 37 markers. Either that or restrict your markers to 17 to match the
Trinity figures.

Does anyone have any opinions on the validity of the time estimates
generated by Network.exe or how to establish the parameters used by the
program?



John

This thread: